HYVE SOLUTIONS TO SHOWCASE MODULAR AI INFRASTRUCTURE AT OCP GLOBAL SUMMIT 2026
Hyve CTO, Dr. Dimitris Ziakas, to Present on Open, Modular Rack-Scale AI Infrastructure for the Agentic Era
WHAT: Hyve Solutions, a US-based design and manufacturing partner for compute, storage and networking systems in AI, cloud, and digital infrastructure, and a wholly owned subsidiary of TD SYNNEX Corporation (NYSE: SNX), will showcase its latest modular AI and data center infrastructure at the 2026 OCP Global Summit.
WHEN: October 12-15, 2026
WHERE: Booth C31, San Jose Convention Center, 150 W. San Carlos Street, San Jose, CA 95113
HYVE’S MODULAR RACK-SCALE SYSTEM ARCHITECTURE
Hyve's modular architecture is the foundation of a long-term engineering partnership with its customers. From Hyve-designed motherboards to systems and fully integrated racks, Hyve engineers work with customers to build infrastructure that evolves with their roadmap, adopting new processors, accelerators and fabrics without a full system redesign. The architecture spans compute, accelerated compute, storage, networking, management, power and cooling.
The booth features an air-cooled scalable AI rack and a liquid-cooled networking rack with advanced optical connectivity. Together they demonstrate four capabilities:
- Accelerated time to market: Common building blocks replace clean-sheet design, shortening engineering, integration and qualification cycles.
- Workload-driven customization: Compute, accelerator, storage and networking elements are configured around each workload on a common platform.
- Rapid technology adoption: New CPUs, GPUs, accelerators, fabrics, cooling and power integrate as the ecosystem advances.
- Integrated scale-up and scale-out: Hyve Solutions Apex™ MGX 6U AI Servers scale up through a shared PCIe fabric built with Astera Labs technology, and Hyve Solutions NexLink™ Ethernet switching scales out across racks.
Hyve's new interactive digital configurator lets attendees build an MGX-based system around their own workload and deployment requirements, then review it with Hyve engineers. It shows how Hyve's modular architecture maps to specific requirements and gives attendees a configuration to build on.
EXECUTIVE SPEAKER SESSION
Open, Modular Rack-Scale AI Infrastructure for the Agentic Era
Presenter: Dr. Dimitris Ziakas, CTO & VP, Architecture, Hyve Solutions
WHEN: Monday, October 12, at 2:30 p.m. PT
WHERE: San Jose Convention Center, Lower Level, Room LL20BDC
As AI infrastructure moves from individual servers toward rack- and multi-rack-scale systems, the architecture challenge spans compute, accelerators, memory, storage, networking, power, cooling and management.
Dr. Ziakas will examine how open standards and modular, interoperable architectures enable this transition while preserving customer choice and accelerating deployment. He will cover composable compute building blocks, scale-up and scale-out fabrics, and open networking, and show how they come together as a cohesive rack-scale system for training, inference and agentic AI.
The question is no longer which processor to deploy, but how to architect the full infrastructure around heterogeneous compute so it can evolve rapidly as AI workloads do.
Media Contact:
Bill Fox
Hughes Agency
For TD SYNNEX Corporation
(864) 271-0718
billf@hughes-agency.com
